✨
International Publishing Platform
Serving Researchers Since 2012

TAT and SLA Analysis for Delivery Process Improvement and SOP Formulation: A Hub-Level Study at Delhivery Limited, Sudhawadi Hub, Pune

DOI : 10.17577/IJERTCONV14IS120016
Download Full-Text PDF Cite this Publication

Text Only Version

TAT and SLA Analysis for Delivery Process Improvement and SOP Formulation: A Hub-Level Study at Delhivery Limited, Sudhawadi Hub, Pune

Vaishnavi Krishnakumar Wani

MBA, School of Business, Indira University, Pune, India 411033

Dr. Nitin Joshi

Associate Professor and Area Chair – MBA, School of Business, Indira University,

Pune, India 411033 nitin.joshi@indirauniversity.edu.in

Maj. Ankan Mondal

Director and External Guide, Delhivery Limited, Sudhawadi Hub, Pune, Maharashtra, India

Dr. Sanjay Asati

Assistant Professor – MBA, School of Business, Indira University,

Pune, India 411033 sanjay.asati@indirauniversity.edu.in

Abstract – Standardised hub operations are central to delivery reliability in Indias express logistics sector, where a shipment passes through several handovers and each can add dwell time, exceptions and Service Level Agreement (SLA) risk. Drawing on Lean thinking, standard work, 5S and PlanDoCheckAct (PDCA) improvement, this applied study examined Turnaround Time (TAT), SLA adherence and exception handling at Delhivery Limiteds Sudhawadi Hub, Pune. The objectives were to map the end-to-end delivery process and stage-wise handoffs, to analyse TAT, SLA and exception patterns to identify delays and root causes, and to formulate department-wise Standard Operating Procedures (SOPs) and control mechanisms. During a four-month on-the-job training (13 April13 August 2026), the author rotated through Inbound, First/Last Mile Success (FLMS), Processing Center (PC), Outbound, Nido and Falcon weight capture, and then worked mainly with the Problem Solving Team (PST). Because official TAT/SLA dashboards were confidential, the quantitative analysis uses a representative academic dataset of about 980,000 shipments; these figures are illustrative and are not official Delhivery performance data.

In the representative dataset, weighted average TAT was 27.9 hours and SLA adherence 91.9%. TAT improved from 31.4 to 24.9 hours (20.7%) and SLA adherence from 88.6% to 94.2% over the study periods. The Processing Center had the longest stage dwell (6.7 hours), followed by FLMS (4.8 hours) and Outbound (4.2 hours). Route/NTC mismatch was the largest exception category (21.8% of 12,600 cases), whereas untraceable (18.6 hours) and damage (14.8 hours) cases took longest to close. Heavy shipments had the lowest SLA (85.6%) and B2B shipments exceeded B2C (94.0% vs 89.8%). It is concluded that remaining service risk lies in processing dwell and exception ageing rather than in any single failure point. SOPs were proposed for Inbound, FLMS, Processing Center, Outbound, Falcon and PST, supported by a daily stage- wise TAT/SLA control board, exception ageing buckets with escalation thresholds, route/pincode verification, pre-dispatch checks, shift-handover controls and periodic SOP audits and PDCA review.

Keywords TAT, SLA, Standard Operating Procedure (SOP), Exception Management, Lean Logistics, Hub Operations, Root Cause Analysis, Delhivery

  1. INTRODUCTION

    Logistics is a process-intensive business in which physical movement and information movement occur together. A parcel may pass through several operational nodes before reaching the customer, and every handover creates an opportunity for delay, misrouting, damage, loss of visibility or rework. The Indian logistics sector spans express parcel, part and full truckload freight, warehousing, cross-border services and logistics software, and its growth is being driven by e-commerce, organised retail and digital supply chains that raise customer expectations for speed, visibility and reliable exception resolution [12], [13]. Policy initiatives such as the National Logistics Policy and PM GatiShakti further emphasise process reform, digitisation and skilling [14].

    Delhivery Limited, founded in 2011 and headquartered in Gurugram, is an integrated logistics provider whose reported FY26 revenue from services was 10,486 crore, with roughly one billion express parcels handled in the year [12]. At this scale, small process variations at a hub produce large aggregate effects. Hub performance depends on inbound receiving, scanning, sorting, processing, weight verification, staging, dispatch and exception closure, and a delay at one point propagates to the stages that follow. TAT and SLA are therefore useful indicators because they connect process execution with customer-facing service outcomes.

    This study focuses on Delhiverys Sudhawadi Hub in Pune. Its problem is that hub-level variations in scanning, sorting, processing, route verification, weight capture, staging and exception closure add dwell time, and accumulated delay exposes shipments to SLA risk. The study does not assume that every SLA miss originates at the hub; it concentrates on controllable hub-level factors and on the interfaces between departments. It follows a three-step approach: (i) map the shipment flow and its handoffs, (ii) analyse TAT, SLA and exception patterns to locate bottlenecks and root causes, and (iii) convert the findings into department-wise SOPs, checkpoints and escalation rules.

  2. LITERATURE REVIEW AND THEORETICAL BACKGROUND

    1. Literature Review

      Womack et al. [1] showed through the Toyota Production System that productivity improves when non-value-adding activities are systematically removed and flow is emphasised, while Imai [2] established Kaizen as a philosophy of small, sustained improvements. Liker [3] described continuous flow, standard work and disciplined problem solving as management principles, and Rother and Shook [4] presented Value Stream Mapping to visualise material and information flow. Baudin [5] adapted Lean to repetitive assembly through standardised work and efficient material handling, and Monden [6] detailed just-in- time and Kanban-based controlled flow. More recent work links standardisation to practical logistics settings: Shafeek et al. [7] offer a roadmap for implementing standardised work, Rizkya et al. [8] show that 5S organisation reduces waste in warehouses, Agyabeng-Mensah et al. [9] report that Lean Six Sigma can reduce process variation in third-party logistics warehousing, and Quiroz Flores et al. [10] and Padilla-Bendezú et al. [11] find that 5S, standardisation, Poka-Yoke and related Lean tools improve warehouse efficiency and reduce variability. These sources are largely conceptual or warehouse-based; this study applies them jointly at a parcel-hub level by combining direct observation, TAT/SLA analysis and SOP formulation, without claiming a universal causal relationship.

    2. Theoretical Background

    The study rests on the principle that process standardisation creates a baseline for measurement and continuous improvement. The framework links observation, process mapping, measurement, gap identification, SOP formulation, implementation, monitoring and review. An SOP defines the sequence, responsibilities and controls for a recurring process, whereas a policy states organisational intent and a work instruction gives task-level guidance; standard work goes further by fixing the sequence and conditions needed for consistent execution [3], [7].

    Hub delays and rework can be viewed as Lean waste: waiting, unnecessary movement, repeated handling, mis-sorting, excess processing and avoidable exception escalation. 5S and visual management, through clear zone identifiers, standard staging positions and visible exception categories, reduce search time and placement ambiguity [8]. The PDCA cycle supports continual improvement: plan the standard, pilot the SOP, check TAT, SLA, exceptions and ageing, then act by correcting the process or revisng the SOP. Exception management requires a defined path of identification, segregation, documentation, investigation, escalation, resolution and closure.

    Performance is expressed through a small KPI set. TAT is the elapsed time between a defined process start and completion; SLA adherence is the number of eligible cases completed within the committed service window divided by all eligible cases, multiplied by 100. Supporting indicators include exception rate, exception closure time, ageing beyond a threshold, scanning accuracy, route verification accuracy and SOP compliance. Averages alone can conceal long-tail delays, so percentile and ageing views are recommended where timestamp data exist.

  3. MATERIALS AND METHODS

    1. Research Objectives

      The study pursues three objectives: (1) to map and evaluate the end-to-end delivery process and the stage-wise handoffs affecting TAT and SLA at Delhiverys Sudhawadi Hub; (2) to analyse representative TAT, SLA and exception data to identify delays, ageing patterns and operational root causes; and (3) to formulate practical department-wise SOPs and control mechanisms that improve TAT and SLA adherence and standardise exception handling.

    2. Research Design and Data Collection

      A descriptive, observational, project-based case-study design was used during the on-the-job training from 13 April to 13 August 2026. The author spent roughly three to four days in each of Inbound, FLMS, Processing Center, Outbound, Nido and Falcon weight capture to follow the shipment from receipt to dispatch, and then spent most of the training with the Problem Solving Team, which handles damage, untraceable, route- mismatch and other exceptions. The observed flow was Inbound

      Primary Sorting FLMS/Secondary Sorting Processing Center Outbound Staging Dispatch/Linehaul Delivery Node. Qualitative evidence came from direct observation, discussions with associates and supervisors, participation in tasks and photographs of the hub environment. Sector and company information came from secondary sources [12][14].

    3. Analytical Framework and Data Integrity

    Stage TAT was calculated as completion time minus start time, and the overall period figure is weighted by shipment volume. Analysis used descriptive statistics, period trends, stage-wise comparison, exception frequency-and-severity (Pareto-oriented) prioritisation, a 4M-plus-information-and- capacity root cause matrix and SOP control-point analysis. SOP drafts followed eight stages: observe the process, map handovers, identify risks, define the standard sequence, add checkpoints and exceptions, define escalation, pilot or validate, and govern through version control.

    Data integrity note: exact Delhivery TAT/SLA dashboard values were not made available because of confidentiality and access restrictions. To avoid presenting invented values as company data, the quantitative analysis uses a representative academic dataset covering 1330 April, May, June, July and 1 13 August 2026 (about 980,000 shipments). It demonstrates the calculation method and improvement logic and must not be read as official performance. Other limitations are that only one hub was studied, exception behaviour varies by lane and customer contract, no confidential cost or employee data were used, and the proposed SOPs are academic drafts that need formal validation by Delhivery before operational use [15].

  4. DATA ANALYSIS AND DISCUSSION

    Profile the Representative Dataset

    TABLE I.

    Parameter

    Value

    Analytical period

    13 April 13 August 2026

    Representative shipment volume

    980,000

    Weighted average TAT

    27.9 hours

    Weighted SLA adherence

    91.9%

    First-period TAT / SLA

    31.4 hours / 88.6%

    epresentative Dataset Profile (13 April13 August 2026)

    Parameter

    Value

    Final-period TAT / SLA

    24.9 hours / 94.2%

    Table I summarises the dataset. Its purpose is to demonstrate how an analyst would evaluate TAT and SLA over the training period, using weighted averages, trends, stage comparison and exception prioritisation. The volumes are not official Delhivery records.

    Analyse Monthly TAT and SLA Trends

    TABLE II.

    onthly TAT and SLA Summary (Representative Data)

    NTC/route mismatch is the largest category and has a moderate closure time (7.2 hours), suggesting that many cases are resolved by verification and correction. Untraceable (18.6 hours) and damage (14.8 hours) cases are slower because they may need physical search, status verification, investigation and escalation. Exceptions should therefore not be prioritised by count alone: a Pareto view shoMuld combine frequency, closure time and SLA impact, since a low-volume category with very long closure can deserve as much attention as a high-volume one that closes quickly.

    Exception

    Cases

    Share

    Avg. Closure (hrs)

    Packaging/handling

    950

    7.5%

    6.4

    Other

    550

    4.4%

    9.0

    Compare Shipment Types

    TABLE V.

    Period

    Shipments

    Avg. TAT

    (hrs)

    SLA

    On-time

    Late

    Apr 13

    30

    112,400

    31.4

    88.6%

    99,586

    12,814

    May 2026

    238,600

    29.8

    90.7%

    216,410

    22,190

    Jun 2026

    252,900

    27.6

    92.1%

    232,921

    19,979

    Jul 2026

    264,300

    26.1

    93.4%

    246,856

    17,444

    Aug 113

    111,800

    24.9

    94.2%

    105,316

    6,484

    Table II shows a steady fall in average TAT from 31.4 hours in the first partial period to 24.9 hours in the last, a reduction of about 20.7%, while SLA adherence rises from 88.6% to 94.2%. In an operational setting this pattern would indicate lower dwell, faster exception resolution or better coordination between stages. The relationship is not strictly linear because SLA depends on the distribution of individual shipment TATs, not only on the average, which is why percentile and ageing measures are recommended.

    Locate Stage-wise Bottlenecks

    TABLE III.

    tage-wise TAT and SLA (Representative Data)

    Process Stage

    Avg. TAT

    (hrs)

    Stage SLA

    Primary Risk

    Inbound

    3.1

    94.1%

    Receipt/scan queue

    FLMS

    4.8

    93.0%

    Route and pincode mismatch

    Processing Center (PC)

    6.7

    91.2%

    Processing backlog

    Outbound

    4.2

    94.4%

    Staging/vehicle readiness

    Dispatch/Linehaul Handover

    3.6

    95.1%

    Cut-off/dispatch delay

    Table III identifies the Processing Center as the primary bottleneck candidate, with the longest stage TAT (.7 hours) and the lowest stage SLA (91.2%). The cause is not assumed to be a single failure; dwell can arise from volume surges, uneven work release, exception accumulation, scanning or system dependencies, or downstream capacity limits. FLMS ranks second (4.8 hours), and route or pincode verification there matters because a sorting error creates a downstream exception that costs more time than the original check. Outbound time is influenced by staging, consolidation and vehicle readiness.

    Prioritise Exceptions by Frequency and Severity

    TABLE IV.

    xception Category Analysis (Representative Data)

    Exception

    Cases

    Share

    Avg. Closure (hrs)

    NTC/route mismatch

    2,750

    21.8%

    7.2

    Damage

    2,100

    16.7%

    14.8

    Untraceable

    1,950

    15.5%

    18.6

    Weight discrepancy

    1,600

    12.7%

    5.9

    Address/pincode issue

    1,450

    11.5%

    8.1

    Scan/manifest issue

    1,250

    9.9%

    4.7

    hipment Type Analysis (Representative Data)

    Shipment Type

    Share

    Avg. TAT (hrs)

    SLA

    B2B

    60.8%

    25.8

    94.0%

    B2C

    32.6%

    29.4

    89.8%

    Heavy

    6.0%

    36.7

    85.6%

    Document

    0.6%

    18.9

    96.2%

    Table V shows that B2B shipments perform better than B2C (94.0% vs 89.8% SLA), and Heavy shipments have the highest TAT (36.7 hours) and the lowest SLA (85.6%). This is plausible because size, handling complexity, routing and vehicle constraints raise dwell. The result supports segmentation: a single SLA target or improvement action should not be applied uniformly to every shipment category.

    Diagnose Root Causes S

    TABLE VI.

    oot Cause Matrix for TAT and SLA Delays

    Category

    Root Cause Example

    Impact on TAT/SLA

    Proposed Control

    Man

    Limited process familiarity during peak workload

    Rework, wrong segregation, slower exception

    closure

    Training, induction checklist, buddy support

    Machine/System

    Scan, weighing or processing equipment

    interruption

    Queue build- up and delayed status completion

    Preventive checks, escalation, backup

    procedure

    Material/Shipment

    Damage, wrong weight, missing identifier, poor

    packaging

    Extra handling and investigation

    Verification checkpoint and exception segregation

    Method

    Variation in

    sequence or unclear handover

    Repeated handling and waiting

    Standard SOP and shift

    handover checklist

    Information

    Incorrect route, pincodEe

    or NTC data

    Mis-sort and downstream correction

    Mandatory verification before transfer

    Capacity

    Uneven workload release versus

    processing capacity

    Processing backlog and ageing

    Hourly workload monitoring and balancing

    Table VI shows that delay has several interacting causes rather than one. High-frequency problems such as route mismatches and scanning errors call for prevention through mandatory checkpoints and training. High-severity damage and

    untraceable cases call for strong exception governance and ageing escalation, and Processing Center backlog calls for workload visibility and capacity balancing.

    Link Findings to SOP Controls

    TABLE VII.

    inding-to-SOP Control Linkage

    Finding

    Relevant SOP

    Control

    High processing dwell

    PC SOP

    Queue monitoring, batch/priority discipline, ageing escalation

    Route mismatch

    FLMS SOP

    Mandatory route/pincode/NTC verification before transfer

    Dispatch readiness delay

    Outbound SOP

    Pre-dispatch checklist and cut-off confirmation

    Weight discrepancy

    Falcon SOP

    Weight capture, label/system verification and recheck

    Damage /

    untraceable

    PST SOP

    Segregation, investigation, escalation and closure

    Repeated small deviations

    All SOPs

    Visual controls, training and compliance audits

    Table VII converts the findings into controls. Each SOP states purpose, scope, responsible role, step sequence, checkpoints, exception handling and escalation, with document ownership, version control and review dates. Cross-functional rotation also showed that end-to-end TAT is shaped by several handoffs rather than one activity, so stage-wise visibility, clear ownership and disciplined handovers are needed alongside the SOPs themselves.

    Plan the Improvement Roadmap

    TABLE VIII.

    ecommended Controls and Expected Outcomes

    Recommendation

    What

    How

    Expected Outcome

    1. Stage-wise TAT/SLA control

    Make dwell and SLA risk visible before breach

    Daily control board, stage KPIs, PC ageing buckets, FLMS verification, outbound pre-

    dispatch checks

    Earlier intervention, lower dwell, higher SLA

    2. Exception and verification control

    Reduce delay from recurring and severe exceptions

    PST ageing matrix, route/NTC verification, Falcon weight recheck, severity-based

    escalation

    Faster closure, less rework, shorter long- tail ageing

    3. SOP

    governance and continuous improvement

    Turn findings into sustained standard

    work

    Controlled SOPs, training, visual management, audits, weekly

    PDCA review

    Higher compliance, lower variation

    Table VIII outlines three connected interventions. The TAT/SLA control board should report average and percentile TAT, SLA adherence, ageing buckets, stage dwell and the top SLA-risk causes, with hourly ageing review in the Processing Center. Falcon weight capture should follow identification, stable reading, system verification, tolerance and recheck, and escalation. Challenges met during the study, namely restricted data access, process complexity, exception variability and confidentiality, were handled through cross-functional rotation, grouping exceptions by type and severity, and disclosing the representative nature of the data.

  5. FINDINGS, CONCLUSIONS AND SUGGESTIONS

  1. Findings

    The representative analysis shows an improving overall trend, with TAT falling from 31.4 to 24.9 hours and SLA adherence rising from 88.6% to 94.2%, but service risk remains

    concentrated in a few places. FThe Processing Center has the

    longest dwell (6.7 hours) and lowest stage SLA (91.2%), followed by FLMS (4.8 hours). Route/NTC mismatch is the most frequent exception (21.8%), while untraceable and damage cases close slowest (18.6 and 14.8 hours). Heavy shipments (85.6% SLA) and B2C shipments (89.8%) lag B2B and document shipments. Process observation confirmed that TAT is a cumulative result of several handoffs.

  2. Conclusions

    The three objectives were achieved. Process mapping through cross-functional rotation established the end-to-end flow and control points; TAT/SLA and exception analysis identified the Processing Center, FLMS routing and slow- closing exceptions as priorities; and the findings were turned into department-wise SOPs and controls. TAT and SLA improvement at the hub should therefore be approached through three connected interventions: control stage-wise handoffs, use TAT/SLA and exception analysis to set ageing and root cause priorities, and institutionalise practical SOPs. As the quantitative figures are representative, official performance decisions should rely on controlled company data.

  3. Suggestions

    To translate the findings into practice, the hub should follow a phased plan: R

    Stage-wise control: Establish a daily TAT/SLA control board with hourly ageing review and workload balancing at the Processing Center, a pre-dispatch readiness checklist at Outbound, and mandatory route/pincode/NTC verification at FLMS.

    Exception management: Run exceptions through the PST with separate response logic for high-frequency (route mismatch) and high-severity (damage, untraceable) cases, using ageing buckets, named ownership and escalation thresholds, and standardise Falcon weight recheck.

    SOP governance: Maintain controlled SOPs for Inbound, FLMS, Processing Center, Outbound, Falcon and PST with ownership, version control, training acknowledgement, visual management, competency checklists and weekly PDCA review.

  4. Implications of the study

    • Hub management: Stage-level visibility and ageing escalation allow earlier intervention before an SLA breach occurs.

    • Frontline associates and supervisors: Standard steps, checkpoints and handover checklists reduce reliance on memory and help new staff learn consistently.

    • PST and quality teams: Exception categories ranked by frequency and closure time give a factual basis for escalation and for tracking recurring causes.

  5. Future Scope of the study

A future study with controlled shipment-level timestamps for receipt, inbound scan, primary sort, FLMS sort, PC completion,

outbound staging, dispatch and delivery, together with shipment type, lane, exception code and closure timestamp, would allow percentile TAT, SLA-breach decomposition and statistical root cause analysis. It could also pilot the proposed SOPs and measure their effect on TAT, SLA adherence, exception closure and process variation, and extend the work to other hubs.

ACKNOWLEDGMENT

I express my sincere gratitude to Dr. Nitin Joshi, Learning Facilitator (Operations and Supply Chain Management), for his valuable guidance and constant encouragement, and to Maj. Ankan Mondal, External Guide and Director, Delhivery Limited, Sudhawadi Hub, Pune, for his esteemed suggestions and support throughout the project.

REFERENCES

  1. J. P. Womack, D. T. Jones, and D. Roos, The Machine That Changed the World: The Story of Lean Production. New York, NY, USA: Rawson Associates, 1990.

  2. M. Imai, Kaizen: The Key to Japans Competitive Success. New York, NY, USA: McGraw-Hill, 1986.

  3. J. K. Liker, The Toyota Way: 14 Management Principles from the Worlds Greatest Manufacturer. New York, NY, USA: McGraw-Hill, 2004.

  4. M. Rother and J. Shook, Learning to See: Value Stream Mapping to Add Value and Eliminate Muda. Cambridge, MA, USA: Lean Enterprise Inst., 1998.

  5. M. Baudin, Lean Assembly: The Nuts and Bolts of Making Assembly Operations Flow. New York, NY, USA: Productivity Press, 2002.

  6. Y. Monden, Toyota Production System: An Integrated Approach to Just- In-Time, 4th ed. Boca Raton, FL, USA: CRC Press, 2012.

  7. H. Shafeek et al., standardized work and Lean implementation, 2018. [Full citation to be completed]

  8. I. Rizkya et al., 5S implementation in warehouse environments, 2021. [Full citation to be completed]

  9. Y. Agyabeng-Mensah et al., Lean Six Sigma in third-party logistics warehousing, 2023. [Full citation to be completed]

  10. J. C. Quiroz Flores et al., Lean warehousing: 5S, Poka-Yoke and TPM, 2023. [Full citation to be completed]

  11. J. Padilla-Bendezú et al., systematic review of Lean warehousing, 2025. [Full citation to be completed]

  12. Delhivery Limited, Annual reports and FY25/FY26 results and corporate disclosures, Gurugram, India, 2026.

  13. India Brand Equity Foundation (IBEF), Indian logistics and express logistics sector reports. [Online]. Available: https://www.ibef.org

  14. Department for Promotion of Industry and Internal Trade (DPIIT), National Logistics Policy and logistics-sector initiatives, Government of India, 2022.

  15. Delhivery Limited, Sudhawadi Hub, Pune, Internship observations, process discussions and photographic evidence, AprilAugust 2026, unpublished, 2026.